7805 needs about 2 volts of clearance to maintain regulation. Data sheet for
the 3 manufacturers I looked up shows it needing at least 0.3uf on the input
to prevent oscillation (which is what is causing the noise on your input
peaks), however you should use a much larger cap on the input to smooth out
the ripple from the bridge as well.

Linear regulators can be picky about their needs for input/output caps. It
varies even more with LDO types. Even the type of cap matters with some of
them...

Chris


"robb" wrote in message
...

"robb" wrote in message
...
yes i googled and found lots of different mods to the basic
5v
*7805* regulator power supply
that is {AC in, bridge, 7805, 5v out}

(images on a.b.s.e)

i built and experimenting with the 5v DC psu
i came across a strange (to me) oscope waveform.

using most basic psu {18vac in, bridge, 7805, 5v out}
with a (470 Ohm + standard LED) power indicator on the 5v side

1st
i oscope the (bridge +) and (7805 out) and i see typical
expected waves... that is bridge + positive 16 v sine peaks
squashed together and 7805 out is 5v DC but with small dips at
intervals cooresponding to positive sine peak dips

2nd (**problem here **)
i add .1uF cap between 7805 out and ground (high frequency
decoupling ) and the bridge + output changes to some high
frequency fuzz on the tops of positive sine peaks.

When i zoom in on the ?fuzz? it is a series of about 1 uSec wide
spikes that are space about 10 uSec apart and the magnitude
follows shape of the sine wave ? (see pic in a.b.s.e) thats the
best this amateur can dexcribe

3rd ( ? problem fix ? )
all of these things by themselves will make fuzz go away....
- add a .1uF cap between the -/+ bridge output
- remove the LED
- add smoothing cap to either side

any ideas to help understand what this fuzz is ?

thanks for any help ,
robb